Characterization of a Bacterial β-1, 3-Galactosyltransferase with

© Copyright 2008 by the American Chemical Society

Volume 47, Number 5

February 5, 2008

Articles Characterization of a Bacterial β-1,3-Galactosyltransferase with Application in the Synthesis of Tumor-Associated T-Antigen Mimics† Wen Yi,‡,| Ramu Sridhar Perali,§ Hironobu Eguchi,‡ Edwin Motari,§ Robert Woodward,§ and Peng George Wang*,‡,§,| Department of Biochemistry, Department of Chemistry, and The Ohio State Biochemistry Program, The Ohio State UniVersity, Columbus, Ohio 43210 ReceiVed October 15, 2007; ReVised Manuscript ReceiVed NoVember 14, 2007 ABSTRACT: T-Antigen (Gal-β1,3-GalNAc-R-O-Ser/Thr) is an important precursor of mucin-type O-glycans. T-Antigen is found to be closely associated with cancer progression and metastasis and has been used to develop carbohydrate-based anticancer vaccines. Enzymatic synthesis of T-antigen disaccharides have relied on the use of β-1,3-galactosyltransferases recently cloned and characterized from several eukaryotic organisms. However, its application is limited by the difficulty of obtaining homogeneous enzymes and the strict substrate specificity of enzymes. Recently, a number of bacteria have been found to express carbohydrate structures that mimic host glycans. The corresponding glycosyltransferases have been exploited in the facile synthesis of a number of clinically important glycoconjugate mimics. In this study, we biochemically characterized a bacterial β-1,3-galactosyltransferase (WbiP) from Escherichia coli O127, which expresses a T-antigen mimic in the lipopolysaccharide (LPS) structure. Substrate study showed that WbiP could readily glycosylate a series of N-acetylgalactosamine (GalNAc) analogues with R-substitutions at the reducing end, including glycosylated Ser and Thr (GalNAc-R-O-Ser/Thr), which illustrates the use of WbiP for the facile synthesis of T-antigens. Alignment of a group of putative bacterial β-1,3-galactosyltransferases revealed the presence of two conserved DXD motifs, possibly suggesting a different functional role of each motif. Site-directed mutagenesis, enzyme kinetics as well as UDP-bead binding assays were carried out to investigate the role of each DXD motif in WbiP. The results suggest that 88DSD90 is critical in the binding of sugar donor UDP-Gal, whereas 174DYD176 may participate in the binding of the sugar acceptor. This study expands the scope of using bacterial glycosyltransferases as tools for in Vitro synthesis of glycoconjugate mimics with clinical significance.

Mucin-type O-linked glycosylation is a ubiquitous protein post-translational modification in higher eukaryotes and is † P. G. Wang acknowledges support from an endowed Ohio Eminent Scholar Professorship on Macromolecular Structure and Function in the Department of Biochemistry at The Ohio State University. * To whom correspondence should be addressed. Phone: (614) 2929884. Fax: (614) 688-3106. E-mail: [email protected]. ‡ Department of Biochemistry. § Department of Chemistry. | The Ohio State Biochemistry Program.

involved in a number of fundamental biological processes (1). For example, mucin-type O-glycans can protect proteins from proteolytic degradation (2), regulate the serum half-life of chemokines in ViVo (3, 4), modulate the intracellular trafficking of proteins (5), mediate cell adhesion events such as sperm-egg fertilization (6), microbe-host interaction, and viral infection (7), and serve as tumor-associated antigens (8). Mucin-type O-glycan is initiated with an R-N-acetylgalactosamine (GalNAc1) attached to the hydroxyl group of

10.1021/bi7020712 CCC: $40.75 © 2008 American Chemical Society Published on Web 01/08/2008

1242 Biochemistry, Vol. 47, No. 5, 2008 Ser/Thr side chains and elaborated by Golgi-resident glycosyltransferases to generate a series of core structures (1). Core-1 structure (also called T-antigen, Gal-β1,3-GalNAcR-O-Ser/Thr) is the precursor of surface antigens in human epithelial and blood group cells (9). T-Antigens are masked in normal tissues but uncovered during malignancy (10). Previous research has shown that T-antigens are expressed in >90% of primary human carcinomas, and high expression levels of T-antigens are correlated with increased metastasis and aggressiveness of several types of cancers (11, 12). Therefore, T-antigens have been used in clinical trials to develop carbohydrate-based cancer vaccines. Various mucosal pathogens express carbohydrate structures on their cell surface that mimic host structures as a means of evading immune clearance. For example, Haemophilus influenzae is a pathogen that routinely colonizes the upper respiratory tract. It is often associated with serious infectious diseases such as meningitis, septicemia, and emphysema. H. influenzae strain Rd express a lipooligosaccharide, globotetraose (13), on its cell surface that can mimic host glycolipids, which is important for the pathogenesis of diseases. Campylobacter jejuni has been recognized as a major cause of acute gastroenteritis in humans. One of the most common strains associated with the development of Guillain-Barre syndrome, C. jejuni O:19 has been shown to display molecular mimicry of gangliosides (GM1, GD1a, GD3, and GT1a) (14). Another common gastric pathogen, Helicobacter pylori, expresses histo blood group antigens, most of which are type II glycoconjugate antigens, Lewis X (Lex) and Lewis Y (Ley) (15-17). The expression of Lewis antigens by H. pylori has also been suggested to play a role in the pathogenesis of chronic type B gastritis and gastric and duodenal ulcers. E. coli O127 belongs to the O serogroup of enteropathogenic Escherichia coli (EPEC) strains, which are important causes of infantile diarrhea in developing countries (1820). E. coli O127 strains have been isolated from children with diarrhea worldwide. In addition, E. coli O127 was reported to possess high human blood group H (O) activity. The elucidation of the chemical structure of the cell surface O-antigen (21) confirmed that E. coli O127 expressed mimicry of human blood group H antigen (Figure 1). The O-antigen biosynthetic gene cluster (GenBank Accession no. AY493508) contains multiple genes needed for the assembly of E. coli O127 polysaccharide structures. Among them, three genes (orf3, orf12, and orf13) encode putative glycosyltransferases involved in the synthesis of repeating oligosaccharide units. Orf12 (WbiP) contains a conserved domain found in glycosyltransferase family 2. It shows 59% protein sequence identity and 82% similarity to WbnJ from E. coli O86, which was previously characterized as a β-1,3-galactosyltransferase (22). Thus we propose that WbiP encodes a β1,3-galactosyltransferase that makes the Gal-β1,3-GalNAc moiety (Tantigen mimicry) present in the repeating unit structure. In 1 Abbreviations: ESI-MS, electrospray ionization mass spectrometry; Gal, galactose; GalNAc, N-acetylgalactosamine; HPLC, high performance liquid chromatography; MALDI-MS, matrix-assisted laser desorption/ionization-mass spectrometry; NMR, nuclear magnetic resonance; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; Ser, serine; Thr, threonine, TLC, thin layer chromatography; UDP-Gal, uridine 5′-diphospho-galactose; UDP-GalNAc, uridine 5′-diphospho-N-acetylgalactosamine; UDP-Glc, uridine 5′diphospho-glucose; GDP-Fuc, guanosine 5′-diphospho-fucose.

Yi et al.

FIGURE 1: O-polysaccharide structure of E. coli O127 and its biosynthetic gene cluster. (T-Antigen mimic structure catalyzed by WbiP is highlighted.)

this study, we present a detailed biochemical characterization of WbiP and a series of mutants. Moreover, we show that this bacterial β-1,3-galactosyltransferase is implicated in the facile synthesis of T-antigen mimics. MATERIALS AND METHODS Bacterial Strains, Plasmids, and Materials. E. coli O127: K63(B8) was obtained from American Type Culture Collection (Rockville, MD). E. coli competent cell DH5R [lacZ∆M15 hsdR recA] was from Gibco-BRL Life Technology. E. coli competent cell BL21 (DE3) [F ompT hsdSB (rBmB) gal dcm (DE3)] were from Novagen Inc. (Madison, WI). Expression plasmid pET28a was purchased from Novagen (Carlsbad, CA). HiTrap Chelating Ni HP column was obtained from GE Healthcare (Piscataway, NJ). All other chemicals and solvents were from Sigma-Aldrich. Cloning, Expression, and Purification of WbiP. The wbiP gene was amplified by PCR from the E. coli O127 chromosome. The primers with restriction sites underlined for amplification of each gene were as follows: Forward: 5′ GAGATATACATATGAAAAATGTTGGTTTTATTG (NdeI); Reverse: 5′ CCGCCTCGAGTCAACCTAAAATAATGCTTTTATATG (XhoI). The DNA fragments obtained were digested with corresponding restriction enzymes and inserted into pET28a vector linearized by the same restriction enzymes to form pET28a-wbiP recombinant plasmid. The recombinant plasmid was confirmed by restriction mapping and sequencing. The correct constructs were subsequently transformed into E. coli BL21 (DE3) for protein expression. E. coli BL21 (DE3) harboring the recombinant plasmid was grown at 37 °C in 1 L of Luria-Bertani (LB) medium with 35 ug/mL kanamycin antibiotic. When the cells were grown to OD 0.5-0.7, isopropyl 1-thio-β-D-galactopyranoside (IPTG) was added to a final concentration of 0.4 mM. Expression was allowed to proceed for 12 h at 25 °C. Cells were harvested, washed with 20 mM Tris-HCl (pH ) 7.0) and stored at -80 °C until needed. In protein purification, the cell pellet was resuspended in buffer (20 mM TrisHCl, pH 7.5, 0.5 M NaCl, 5 mM imidazole, 1 mM PMSF) and disrupted by brief sonication (Branson Sonifier 450, VWR Scientific) on ice. The lysate was cleared by centrifugation (15 000g, 20 min, 4 °C) and loaded at a low rate of 3 mL/min onto a 5 mL Ni2+-NTA chelating column equilibrated with 20 mM Tris-HCl, pH 7.5, 0.5 M NaCl, and 5 mM imidazole. The column was washed with 10

Characterization of a Bacterial β-1,3-Galactosyltransferase

Biochemistry, Vol. 47, No. 5, 2008 1243

column volumes of 50 mM imidazole in the same buffer, and the protein was eluted with 250 mM imidazole, followed by analysis using SDS-PAGE. RadioactiVe Enzymatic Assay. Protein concentration was determined by Bradford method. One unit of enzyme activity was defined as the amount of enzyme required to transform 1 µmol of sugar donor to acceptor per minute at 37 °C. Enzyme assays were performed at 37 °C for 2 h in a final volume of 100 µL containing 20 mM Tris-HCl, pH 7.0, 10 mM MnCl2, 0.3 mM radioactively labeled UDP-D-[6-3H]galactose (specific activity: 4.3 Ci/mmol) as sugar donor, 10 mM monosaccharide N-acetylgalactosamine (GalNAc) as sugar acceptor, and variable amount of enzymes. Acceptor was omitted as the blank control. The reaction was terminated by adding 100 µL of ice cold 0.1M EDTA. Dowex 1 × 8-200 chloride anion exchange resin was then added in a water suspension (0.8 mL, v/v ) 1/1). After centrifugation, supernatant was collected in a 20 mL plastic vial and ScintiVerse BD (10 mL) was added. The vial was vortexed thoroughly before the radioactivity of the mixture was measured in a liquid scintillation counter (Beckmann LS3801 counter). Enzyme Assay with HPLC and MALDI-MS. To a tube containing 0.05 mg of GalNAc were added 10 µL of 0.1 M Tris-HCl, pH 7.5, 5 µL of 0.1 M MnCl2, 5 µL of 50 mM UDP-Gal, and 0.02 mg of WbiP (20 µL). The reaction was run at room temperature for 1 h and quenched by addition of 50 µL of cold methanol. The mixture was centrifuged to remove the precipitated proteins. The supernatant was evaporated to dryness. The mixture was then labeled with 2-aminobenzamide (2-AB) as described. Postlabeling cleanup was done by applying the labeling sample onto a Whatman 3MM chromatography paper and performing ascending chromatography in acetonitrile for 2 h. The labeling product was eluted from the paper with 0.5 mL of H2O and further separated on a normal phase analytic HPLC column as described. The peaks were detected with a fluorescence detector (λex ) 330 nm, λem ) 420 nm) and collected and characterized with MALDI-MS. Site-Directed Mutagenesis of WbiP. QuikChange SiteDirected Mutagenesis Kit (Stratagene) was used to construct all the mutants from the parent plasmid, pET28a-WbiP, using the primers shown in the Supporting Information Table S1. Expression, purification, and enzyme activity assay were carried out following the abovementioned procedures. Kinetic Studies of Wild-Type and Mutants. Reactions were performed at 37 °C for 40 min in a reaction buffer containing 20 mM Tris-HCl (pH 7.5), 10 mM MnCl2, 0.3 mM UDPD-[6-3H]-Gal and 20 µg of enzyme. To determine apparent Km values, the concentration of GalNAc was varied as follows: 10 mM, 20 mM, 40 mM, and 80 mM. To determine the Km value for UDP-Gal, 3 µM of UDP-D-[6-3H]-Gal was supplemented with different amounts of cold UDP-Gal to achieve varying final concentrations (0.08, 0.1, 0.3, and 0.6 mM), with a fixed GalNAc concentration at 20 mM. The parameters Km and Vmax were obtained by Lineweaver-Burk plots of substrate concentration-initial velocity. Enzymatic Synthesis of T-Antigen Mimics and Structural Characterization. A one-pot reaction was conducted for 2 days at room temperature in a final volume of 1.0 mL containing 20 mM Tris-HCl (pH 7.5), 10 mM MnCl2, 10 mM UDP-Gal, 15 mM acceptor GalNAcR-OMe, and 20 µg

of WbiP. The progress of the reaction was monitored by thinlayer chromatography [i-PrOH/H2O/NH4OH ) 8:2:2 (v/v/ v)] conducted on Baker Si250F silica gel TLC plates. Products were visualized by staining solution (anisaldehyde/ MeOH/H2SO4 ) 1:15:2 (v/v/v)). After complete conversion of donor substrate, protein was removed by brief boiling and centrifugation (12 000g, 5 min). The supernatant mixture was purified by gel filtration chromatography Bio-Gel P2 (BioRad, CA). The desired fractions were pooled, lyophilized, and stored at -20 °C. Mass spectra (ESI) were run in the negative mode at the mass spectrometry facility at the Ohio State University. Product structure was identified by 1H and 13C NMR spectroscopy using a 500-MHz Varian VXR500 NMR spectrometer. Product structure was identified through onedimensional (selective COSY, relay COSY, and NOE) and two-dimensional (COSY, HMQC, NOESY, and HMBC) 1H/ 13C NMR. The oligosaccharide product was repeatedly dissolved in D2O and lyophilized before NMR spectra were recorded at 303 K in a 5 mm tube. 1H NMR (500 MHz, D2O): 1.98 (s, 3H), 3.35 (s, 3H), 3.47 (dd, 1H, J ) 9.9 Hz, J ) 7.8 Hz), 3.57 (dd, 1H, J ) 9.9 Hz, J ) 3.4 Hz), 3.61 (dd, 1H, J ) 7.7 Hz, J ) 4.6 Hz), 3.71 (m, 2H), 3.72 (m, 2H), 3.86 (d, 1H, J ) 3.3 Hz), 3.91 (dd, 1H, J ) 7.3 Hz, J ) 5.0 Hz), 3.96 (dd, 1H, J ) 11.1 Hz, J ) 3.1 Hz), 4.18 (d, 1H, J ) 2.9 Hz), 4.29 (dd, 1H, J ) 11.1 Hz, J ) 3.7 Hz), 4.41 (d, 1H, J ) 7.8 Hz), 4.74 (d, 1H, J ) 3.7 Hz). δ 13C NMR (125 MHz, D2O): δ 175.0, 105.1, 98.7, 77.7, 75.3, 72.9, 71.0, 70.8, 69.1, 69.0, 61.6, 61.3, 55.5, 49.0, 22.4. HRMS, m/z: Calculated C15H27NO11 + Na: 420.1476; found: 420.1474. UDP-Binding Assay. A 30 µL aliquot of fresh UDP-beads (CalBiochem) was washed three times with binding buffer (20 mM Tris-HCl pH 7.0, 0.1 M NaCl, 5.0 mM MgCl2). The beads were incubated on a roller at 4 °C for 1 h with 40 µL of purified recombinant wild type WbiP and its variant mutants (50 µg of proteins) in binding buffer. The negative control involved incubation of wild type WbiP with 10 mM UDP to inhibit binding to UDP-beads. The beads were harvested by centrifugation for 1 min at 1000g, followed by washing with binding buffer three times. The bound proteins were treated in SDS loading buffer at 100 °C for 5 min before analysis by SDS-PAGE and western blotting with anti-His antibody. RESULTS Expression, Purification, and Enzymatic Assay of WbiP Protein. The full open reading frame of wbiP gene (750 bp) was amplified from E. coli O127 genomic DNA and subsequently cloned into pET-28a vector. The recombinant plasmid was transformed into E. coli BL21(DE3) strain for induced expression with 0.4 mM IPTG at 25 °C. WbiP protein has an apparent molecular weight of 28 kDa estimated by SDS-PAGE (data not shown), similar to the theoretical value (28.6 kDa) calculated from its predicted amino acid sequence. Primary sequence analysis shows that WbiP has no transmembrane segments, consistent with the result that we were able to purify WbiP without difficulty even though detergents were omitted. To investigate the function of WbiP, we tested a panel of radiolabeled sugar donors (UDP-Gal, UDP-GalNAc, GDP-

1244 Biochemistry, Vol. 47, No. 5, 2008

Yi et al.

FIGURE 2: WbiP-catalyzed reaction and HPLC identification of product.

Fuc, and UDP-Glc) with GalNAc as an acceptor. The results showed that WbiP displayed high activity with UDP-Gal as the sugar donor, nearly 10 fold reduced efficiency with UDPGalNAc (see Supporting Information, Figure S1), and no activity with GDP-Fuc and UDP-Glc. Thus, the donor specificity demonstrated that wbiP encodes a galactosyltransferase, consistent with the bioinformatics analysis. To confirm the galactosyltransferase activity of WbiP, we set up a 50 µL scale reaction, which contained 0.05 mg of GalNAc, 5 µL of 50 mM cold UDP-Gal, and 0.02 mg of WbiP (20 µL). The reaction product was then labeled with 2-aminobenzamide (2-AB) and analyzed via a normal phase analytical HPLC columns. The molecular weight was characterized by MALDI-MS. The 2-AB-labeled starting material GalNAc has a retention volume of 6 mL. When the labeled product of the reaction was analyzed, a new peak with a retention volume of 10 mL was observed (Figure 2). When this peak was collected and subjected to MALDI-MS analysis, a m/z ratio of 526.20 [M + Na]+ was observed (see Supporting Information), consistent with a 2-AB-labeled Gal-GalNAc disaccharide. Therefore, the results from both the radioactive assay and the HPLC/MS assay confirm that WbiP encodes a galactosyltransferase. Characterization of Linkage Specificity and Enzymatic Synthesis of T-Antigen Mimics. To confirm that WbiP catalyzes the formation of the β-1,3-linkage, a reaction was carried out with GalNAc-R-OMe as an acceptor. GalNAc-

R-OMe was used instead of GalNAc because GalNAc-ROMe has the same reactivity as GalNAc (see below) while the fixed configuration of the anomeric carbon by OMe substitution facilitates the NMR analysis. A milligram scale reaction was conducted using 8 mg of GalNAc-OMe and 1.5 equiv of UDP-Gal. The reaction was incubated at room temperature overnight and monitored by TLC. The disaccharide product was isolated by gel filtration chromatography. A total of 11 mg of disaccharide (85% yield) was collected to allow structural analysis by NMR spectroscopy. Substrate Specificity of WbiP. Purified WbiP was used to investigate its specificity toward various sugar acceptors (Table 1). The results indicate that the R-configuration at the reducing end of GalNAc is important for enzyme activity since changing the configuration from R to β makes GalNAc a poor substrate for WbiP (GalNAc-β-OMe has only 15% activity compared to GalNAc-R-OMe). The N-acetyl at C2 of GalNAc is essential for enzyme recognition, since no detectable activity was observed when Gal-R-OMe was used as an acceptor. Furthermore, WbiP can tolerate substitution on the 1-O position of GalNAc, demonstrated by observation that a series of R-substitutions (R-OMe, R-OBn, R-O-Ser/ Thr, and R1,3-GalNAc-PP-C11) have comparable reactivities. It is worth noting that GalNAc-R-O-Ser/Thr are the biosynthetic precursors for T-antigens of mucin-type O-glycans. The fact that these two glycosylated amino acids serve as good substrates indicates that WbiP can be potentially used

Characterization of a Bacterial β-1,3-Galactosyltransferase

Biochemistry, Vol. 47, No. 5, 2008 1245

Table 1: Substrate Specificity of WbiP

FIGURE 3: Metal ion requirement of WbiP. The concentration of EDTA and metal ions are all at 10 mM.

for the large scale synthesis of T-antigens for biomedical studies. Not surprisingly, WbiP readily accepts R-GalNAcR1,3-GalNAc-PP-C11 as a substrate, which is an analogue of the immediate synthetic substrate of WbiP based on the E. coli O127 O-repeating unit structure. Thus taken together, the acceptor pattern shown in the assay is in agreement with the proposed function of WbiP in the biosynthesis of the E. coli O127 O-repeating unit. Determination of Metal Ion Effect on WbiP. Many glycosyltransferases are found to contain a short conserved amino acid sequence called the DXD motif and exhibit a requirement for a divalent metal cation for activity. Sequence analysis shows that WbiP contains two DXD motifs (88DSD90 and 174DYD176), suggesting that WbiP activity requires a metal ion cofactor. To further investigate the metal ion requirements, various divalent metal cations as well as EDTA were incubated in a series of separate reactions. As predicted, WbiP activity requires metal ions with Mn2+ and Mg2+ as the preferred cations (Figure 3). The activity was completely abolished in the absence of metal ions or in the presence of 10 mM EDTA. Thus WbiP belongs to glycosyltransferase superfamily A (GTA), in which the DXD motif plays a role in coordinating the binding of sugar nucleotide donors in the active site.

Site-Directed Mutagenesis Study of DXD Motifs and Enzyme Kinetics of WbiP and Its Mutants. Sequence alignment of a group of bacterial β1,3-galactosyltransferases (putative and characterized) shows that they all possess two DXD or DXD-like motifs (Figure 4), suggesting this might reflect a common catalytic mechanism in these galactosyltransferases. In order to investigate the functional role of each DXD motif, we mutated the aspartic acid residues of the DXD motifs in the WbiP sequence. Aspartic acid residues in each DXD motif were mutated to alanine and glutamic acid, respectively. A total of eight mutations were generated: D88A, D88E, D90A, D90E, D174A, D174E, D176A, and D176E. The mutagenic primer pairs are listed in Supporting Information. WbiP mutants were expressed at a similar level compared to the wild-type (data not shown), suggesting that the mutations in the DXD motifs do not compromise protein folding and stability. The enzyme activity of each mutant was assessed using the radioactive assay. The result (Figure 5) shows that mutating aspartic acid to alanine at 88DSD90 completely abolishes the enzyme activity, while the same mutation at 174DYD176 still retains some marginal activity (12% and 10% for D174A and D176A, respectively). This observation indicates that 88DSD90 may play a more essential role in catalysis than 174DYD176. The change of aspartic acid residue to its conserved counterpart, glutamic acid, in 88DSD90 partially restored enzyme activity (8% for D88E and 10% for D90E), indicating the importance of charge in these positions for catalysis. The same mutations at 174DYD176 restored enzyme activity to a higher level (23% for D174E and 31% for D176E), further confirming that the requirement of aspartic acid residues in this position is less stringent. Kinetic parameters (Table 2) were determined for both the wild-type and mutant enzymes. The wild-type enzyme has apparent Km values of 31 µM and 16.1 mM for donor UDP-Gal and acceptor GalNAc, respectively. These are

1246 Biochemistry, Vol. 47, No. 5, 2008

Yi et al.

FIGURE 4: Amino acid sequence alignment of segments from different β1,3-galactosyltransferases. (Conserved DXD motifs are bracketed with red.) WbiP: from E. coli O127O127:K63(B8), AAR90893; WbnJ: from E. coli O86:H2, AAV80758; WbgO: from E. coli O55:H7, AAL67559; WbdO: from Salmonella enterica subsp. salamae seroVar Greenside 50, AAV34525; WbsK: E. coli O128:B12, AAO37699; WblT: from Photorhabdus luminescens subsp. laumondii TTO1, CAE17191; WbsT: from E. coli O126, DQ465248.

FIGURE 6: UDP-beads binding assay.

D90A mutants have remarkably reduced binding, suggesting that D88 and D90 are involved in the binding of sugar donor UDP-Gal. These results are consistent with the kinetic studies.

FIGURE 5: Enzymatic activity of WbiP mutants. Table 2: Kinetic Parameters of WbiP and Mutants UDP-Gal

GalNAc

WbiP

Kcat (s-1)

Km (µM)

Kcat/Km (s-1 µM-1)

Kcat (s-1)

Km (mM)

Kcat/Km (s-1 mM-1)

WT D88E D90E D174A D174E D176A D176E

11.2 0.08 0.2 2.9 3.7 1.0 3.9

31 232 251 58 47 67 43

0.36 0.00034 0.00079 0.05 0.079 0.015 0.091

15.3 2.2 0.6 0.4 1.7 1.2 5.1

16.1 25.2 32.3 375 175 237 140

0.95 0.087 0.018 0.0011 0.0097 0.0051 0.036

similar to the values obtained in other well-characterized glycosyltransferases. Notably, D88E and D90E mutants have considerably higher Km values for UDP-Gal but quite similar Km values for GalNAc. On the contrary, mutations in D174 and D176 resulted in higher Km values for GalNAc but similar values for UDP-Gal. This observation suggests that 88DSD90 is critical in the binding of sugar donor UDP-Gal, while 174DYD176 may participate in the binding of the sugar acceptor. These two DXD motifs thus have different functional roles in catalysis. UDP-Beads Binding Assay. To further investigate the involvement of the 88DSD90 motif in the binding of the sugar donor, we carried out a UDP-beads binding assay with purified wild-type and mutant enzymes (Figure 6). Wildtype WbiP binds efficiently to UDP-beads. The binding was substantially decreased in the presence of 10 mM UDP, indicating that the binding of WbiP to the beads is specific. D174A and D176A mutants have similar binding efficiency as compared to the wild-type. On the other hand, D88A and

DISCUSSION Cell surface glycoconjugates contain rich structural information that is central in virtually every aspect of biological processes, ranging from protein folding, cellular development, and host-pathogen interaction to immune regulation (23). In recent years, the study of the structure-function relationship of glycoconjugates has been aided tremendously by the in Vitro synthesis of structurally defined glycoconjugates and their analogues. Chemical glycosylation, well documented for its synthetic flexibility, has been successful in preparing a wide range of carbohydrate structures. However, the chemical approach still suffers from tedious protection/deprotection steps and difficult purification of products. On the other hand, with the rapid development of genomic sequencing, a growing number of glycosyltransferases have been characterized with a wide spectrum of acceptor specificity that has proven invaluable for the facile synthesis of a variety of glycoconjugate structures (24). It is found that mammalian glycosyltransferases are relatively hard to express in large quantities and are more rigid in terms of substrate specificity. In contrast, their bacterial counterparts are more easily overexpressed as soluble and active forms without complicated gene manipulation techniques. Furthermore, bacterial glycosyltransferases seem to have broader acceptor-substrate specificities, thereby offering tremendous advantages in the enzymatic synthesis of oligosaccharides and their analogues. In addition, various bacteria exhibit structural mimicry of mammalian glycocon-

Characterization of a Bacterial β-1,3-Galactosyltransferase

Biochemistry, Vol. 47, No. 5, 2008 1247

jugates on their cell surface as a survival and infection strategy. Therefore, the corresponding glycosyltransferases might be explored for the in Vitro synthesis of various glycoconjugates. In fact, a number of bacterial glycosyltransferases have been successfully applied in the synthesis of oligosaccharide moieties of various glycolipids, such as gangliosides (25), blood group antigens (22), and globo-series glycoceramides (13, 25-27). T-Antigen has been recognized as an important tumor antigen that is exploited for the clinical development of carbohydrate-based anticancer vaccines. Various chemical strategies have been developed to synthesize T-antigencontaining glycopeptides (28). On the other hand, several eukaryotic glycosyltransferases have been identified to make T-antigens in different species. Ju et al. (29) cloned and identified a human core-1 β1,3-galactosyltransferase and showed that the enzyme can generate the core-1 structure from a synthetic glycopeptide with GalNAc linked to a Thr residue. Subsequently, a rat core-1 β1,3-galactosyltransferase was cloned (30), characterized, and shown to glycosylate GalNAc-R-O-Ph to provide T-antigen mimicry. More recently, the corresponding enzymes in Drosophilia and C. elegans (31, 32) were also reported. Several bacterial β1,3galactosyltransferases have also been studied, all of which have endogenous acceptor specificity, different from the GalNAcR structure present in the T-antigen. We previously reported identification of a β1,3-galactosyltransferase, WbnJ, involved in the E. coli O86 O-polysaccharide biosynthesis and showed that WbnJ could synthesize a T-antigen analogue (22). However, the detailed biochemical and substrate specificity study has not been carried out. Nonetheless, it was to our knowledge the first bacterial glycosyltransferase to make a T-antigen structure. In our current study, we characterized a second β1,3-galactosyltransferase WbiP from E. coli O127. A detailed substrate specificity study showed that WbiP can be used for the facile synthesis of a variety of T-antigen mimics. Most of the glycosyltransferases in the GTA family possess one or more DXD motifs, which have been shown to be involved in the coordination of a divalent metal ion that is required for the binding of sugar nucleotides in the active site (33, 34). Thus, characterization of the DXD motif is an important step to understanding glycosyltransferase mechanisms. Recently, both structural and biochemical studies revealed that if more than one DXD motif is present, not all the DXD motifs are functionally equivalent. A study (35) in the analysis of two DXD motifs in human xylosyltransferase I showed that the 314DED316 motif did not participate in catalysis, while the 745DWD747 motif was clearly involved in the binding of the sugar nucleotide. A well studied galactosyltransferase LgtC from Neisseria meningitides contains four DXD motifs (36). The crystal structure of LgtC revealed that only two DXD motifs were located within the active site. However, of these two, one of them indeed acts to coordinate the Mn2+ ion with the phosphate group of UDP, whereas the other is involved in the binding of the sugar acceptor. In our study, WbiP possesses two DXD motifs, 88DSD90 and 174DYD176. Mutation studies on these two motifs indicated that they may play different roles in enzyme catalysis. Mutation of aspartic acid into alanine at the 88DSD90 position completely abolished the enzyme activity, while marginal activity was still retained with the same mutations at 174DYD176 position, suggesting

that 88DSD90 is more essential and may directly involve in the binding of sugar nucleotides in the active site. In the kinetic studies, substantially reduced affinity for sugar donor UDP-Gal was observed for mutants D88E and D90E, while the affinity for sugar acceptors was similar to the wild-type. This observation further supports the role of 88DSD90 in the coordination of UDP-Gal binding. On the other hand, similar sugar donor affinity but reduced sugar acceptor affinity were observed for the mutants of 174DYD176, suggesting that the 174 DYD176 motif may be involved in the binding of the sugar acceptor. A third biochemical study, UDP-beads binding assay, adds another piece of experimental evidence to the differential roles of these two DXD motifs, in that reduced binding efficiency was observed in 88DSD90 mutants but not in 174DYD176 mutants. It is worth noting that a number of bacterial β1,3-galactosyltransferases annotated in the database appear to possess these two conserved DXD motifs, suggesting a common glycosyl transfer mechanism. The elucidation of the functional roles of these two DXD motifs requires determination of the enzyme 3-D structure. ACKNOWLEDGMENT We thank Juliana Pernik for her excellent administrative assistance. SUPPORTING INFORMATION AVAILABLE Sequences of primers used in site-directed mutagenesis of WbiP, donor specificity study of WbiP, and MS spectrum of 2-AB labeled product of WbiP-catalyzed reaction with GalNAc as the substrate. This material is available free of charge via the Internet at http://pubs.acs.org. REFERENCES 1. Hang, H. C., and Bertozzi, C. R. (2005) The chemistry and biology of mucin-type O-linked glycosylation, Bioorg. Med. Chem. 13 (17), 5021-5034. 2. Garner, B., et al. (2001) Structural elucidation of the N- and O-glycans of human apolipoprotein(a): role of o-glycans in conferring protease resistance, J. Biol. Chem. 276 (25), 2200022208. 3. Ellies, L. G., et al. (2002) Sialyltransferase ST3Gal-IV operates as a dominant modifier of hemostasis by concealing asialoglycoprotein receptor ligands, Proc. Natl. Acad. Sci. U.S.A. 99 (15), 10042-10047. 4. Smith, K. A. (1988) Interleukin-2: inception, impact, and implications, Science 240 (4856), 1169-1176. 5. Altschuler, Y., et al. (2000) Clathrin-mediated endocytosis of MUC1 is modulated by its glycosylation state, Mol. Biol. Cell 11 (3), 819-831. 6. Primakoff, P., and Myles, G. D. (2002) Penetration, adhesion, and fusion in mammalian sperm-egg interaction, Science 296 (5576), 2183-2185. 7. Hooper, L. V., and Gordon, J. I. (2001) Commensal host-bacterial relationships in the gut, Science 292 (5519), 1115-1118. 8. Brockhausen, I. (1999) Pathways of O-glycan biosynthesis in cancer cells, Biochim. Biophys. Acta 1473 (1), 67-95. 9. Sotiriadis, J., et al. (2004) Thomsen-Friedenreich (T) antigen expression increases sensitivity of natural killer cell lysis of cancer cells, Int. J. Cancer 111 (3), 388-397. 10. Springer, G. F. (1997) Immunoreactive T and Tn epitopes in cancer diagnosis, prognosis, and immunotherapy, J. Mol. Med. 75 (8), 594-602. 11. Coon, J. S.,Weinstein, R. S., and Summers, J. L. (1982) Blood group precursor T-antigen expression in human urinary bladder carcinoma, Am. J. Clin. Pathol. 77 (6), 692-699. 12. Laurent, J. C., Noel, P., and Faucon, M. (1978) Expression of a cryptic cell surface antigen in primary cell cultures from human breast cancer, Biomedicine 29 (8), 260-261.

1248 Biochemistry, Vol. 47, No. 5, 2008 13. Shao, J., et al. (2002) Overexpression and biochemical characterization of beta-1,3-N-acetylgalactosaminyltransferase LgtD from Haemophilus influenzae strain Rd, Biochem. Biophys. Res. Commun. 295 (1), 1-8. 14. Gilbert, M., et al. (2000) Biosynthesis of ganglioside mimics in Campylobacter jejuni OH4384. Identification of the glycosyltransferase genes, enzymatic synthesis of model compounds, and characterization of nanomole amounts by 600-MHz H and C NMR analysis, J. Biol. Chem. 275 (6), 3896-3906. 15. Wang, G., et al. (1999) Novel Helicobacter pylori alpha 1,2fucosyltransferase, a key enzyme in the synthesis of Lewis antigens. Microbiology (Reading, U.K.) 145 (11), 32453253. 16. Ge, Z., et al. (1997) Cloning and heterologous expression of an alpha 1,3-fucosyltransferase gene from the gastric pathogen Helicobacter pylori, J. Biol. Chem. 272 (34), 2135721363. 17. Chan, N. W. C., et al. (1995) The biosynthesis of Lewis X in Helicobacter pylori, Glycobiology 5 (7), 683-688. 18. Rivas, M., et al. (2000) Intestinal bleeding and occlusion associated with Shiga toxin-producing Escherichia coli O127:H21. Instituto Nacional de Enfermedades Infecciosas, An. Dr. Carlos G. Malbran, Buenos Aires, Argentina, 249-252. 19. Barua, D. (1962) An outbreak of infantile diarrhoea due to Escherichia coli O127:B8, Indian J. Med. Res. 50, 612. 20. Stulberg, C. S., et al. (1955) Escherichia coli O127:B8, a pathogenic strain causing infantile diarrhea. I. Epidemiology and bacteriology of a prolonged outbreak in a premature nursery, Am. J. Dis. Child. 90 (2), 125-134. 21. Widmalm, G., and Leontein, K. (1993) Structural studies of the Escherichia coli O127 O-antigen polysaccharide, Carbohydr. Res. 247, 255-262. 22. Yi, W., et al. (2005) Escherichia coli O86 O-antigen biosynthetic gene cluster and stepwise enzymatic synthesis of human blood group B antigen tetrasaccharide, J. Am. Chem. Soc. 127 (7), 20402041. 23. Feizi, T., and Bundle, D. (1994) Carbohydrates and glycoconjugates. Editorial overview, Curr. Opin. Struct. Biol. 4 (5), 673676. 24. Blixt, O., and Razi, N. (2006) Chemoenzymatic synthesis of glycan libraries, Methods Enzymol 415, 137-153.

Yi et al. 25. Antoine, T., et al. (2003) Large-scale in vivo synthesis of the carbohydrate moieties of gangliosides GM1 and GM2 by metabolically engineered Escherichia coli, ChemBioChem 4 (5), 406412. 26. Chen, X., et al. (2002) Reassembled biosynthetic pathway for largescale carbohydrate synthesis: alpha-gal epitope producing “superbug”, ChemBioChem 3 (1), 47-53. 27. Randriantsoa, M., et al. (2007) Synthesis of globopentaose using a novel beta 1,3-galactosyltransferase activity of the Haemophilus influenzae beta 1,3-N-acetylgalactosaminyltransferase LgtD, FEBS Lett. 581 (14), 2652-2656. 28. Brocke, C., and Kunz, H. (2002) Synthesis of tumor-associated glycopeptide antigens, Bioorg. Med. Chem. 10 (10), 3085-3112. 29. Ju, T., et al. (2002) Cloning and expression of human core 1 beta 1,3-galactosyltransferase, J. Biol. Chem. 277 (1), 178-186. 30. Ju, T., Cummings, R. D., and Canfield, W. M. (2002) Purification, characterization, and subunit structure of rat core 1 beta 1,3galactosyltransferase, J. Biol. Chem. 277 (1), 169-177. 31. Ju, T., Zheng, Q., and Cummings, R. D. (2006) Identification of core 1 O-glycan T-synthase from Caenorhabditis elegans, Glycobiology 16 (10), 947-958. 32. Muller, R., et al. (2005) Characterization of mucin-type core-1 beta 1-3 galactosyltransferase homologous enzymes in Drosophila melanogaster, FEBS J. 272 (17), 4295-4305. 33. Davies, G. J., Gloster, T. M., and Henrissat, B. (2005) Recent structural insights into the expanding world of carbohydrate-active enzymes, Curr. Opin. Struct. Biol. 15 (6), 637-645. 34. Schuman, B., Alfaro, J. A., and Evans, S. V. (2007) Glycosyltransferase structure and function, Top. Curr. Chem. 272 (Bioactive Conformation I), 217-257. 35. Goetting, C., et al. (2004) Analysis of the DXD Motifs in Human Xylosyltransferase I Required for Enzyme Activity, J. Biol. Chem. 279 (41), 42566-42573. 36. Persson, K., et al. (2001) Crystal structure of the retaining galactosyltransferase LgtC from Neisseria meningitidis in complex with donor and acceptor sugar analogs, Nat. Struct. Biol. 8 (2), 166-175. BI7020712